As an economist, I am often frustrated by the balancing act in economics. On one side, we have to make assumptions about human behavior in order to apply math to it and perform theoretical and empirical analysis. On the other side, human behavior is not mathmatical, so we have to be able to think, talk and theorize about behavior in non mathmatical ways.
But modern economic science rejects the second part. The most common assumptions to make us humans mathmatically tractable are that we are perfectly self interested and have perfect information. Both of which are NEVER true. So most economic science is about situations that in fact never occur, and we have to cautious about how we interpret it.
Which is why I found Will Wilkinson's recent musing on human self interest and altruism and cooperations to be food for thought.